Upload folder using huggingface_hub

m08/benchmark_results.json

@@ -0,0 +1,540 @@
+{
+  "model_id": "m08",
+  "checkpoint": "results/m08/final.pt",
+  "trained_step": 15258,
+  "num_fewshot": 0,
+  "limit": null,
+  "results": {
+    "lambada_openai": {
+      "name": "lambada_openai",
+      "alias": "lambada_openai",
+      "sample_len": 5153,
+      "perplexity,none": 660.9984145594016,
+      "perplexity_stderr,none": 35.57789788830106,
+      "acc,none": 0.172132738210751,
+      "acc_stderr,none": 0.005259254241062246
+    },
+    "hellaswag": {
+      "name": "hellaswag",
+      "alias": "hellaswag",
+      "sample_len": 10042,
+      "acc,none": 0.27394941246763593,
+      "acc_stderr,none": 0.004450718673552817,
+      "acc_norm,none": 0.277733519219279,
+      "acc_norm_stderr,none": 0.004469659042824522
+    },
+    "piqa": {
+      "name": "piqa",
+      "alias": "piqa",
+      "sample_len": 1838,
+      "acc,none": 0.5990206746463548,
+      "acc_stderr,none": 0.011434766962108364,
+      "acc_norm,none": 0.5805223068552775,
+      "acc_norm_stderr,none": 0.011513551649399906
+    },
+    "sciq": {
+      "name": "sciq",
+      "alias": "sciq",
+      "sample_len": 1000,
+      "acc,none": 0.664,
+      "acc_stderr,none": 0.014944140233794895,
+      "acc_norm,none": 0.585,
+      "acc_norm_stderr,none": 0.015589035185604594
+    },
+    "arc_easy": {
+      "name": "arc_easy",
+      "alias": "arc_easy",
+      "sample_len": 2376,
+      "acc,none": 0.4574915824915825,
+      "acc_stderr,none": 0.010222638127749506,
+      "acc_norm,none": 0.4116161616161616,
+      "acc_norm_stderr,none": 0.010098218646714813
+    },
+    "winogrande": {
+      "name": "winogrande",
+      "alias": "winogrande",
+      "sample_len": 1267,
+      "acc,none": 0.5288082083662194,
+      "acc_stderr,none": 0.014029141615909664
+    },
+    "blimp_adjunct_island": {
+      "name": "blimp_adjunct_island",
+      "alias": "blimp_adjunct_island",
+      "sample_len": 1000,
+      "acc,none": 0.848,
+      "acc_stderr,none": 0.01135891830347525
+    },
+    "blimp_anaphor_gender_agreement": {
+      "name": "blimp_anaphor_gender_agreement",
+      "alias": "blimp_anaphor_gender_agreement",
+      "sample_len": 1000,
+      "acc,none": 0.856,
+      "acc_stderr,none": 0.01110798754893916
+    },
+    "blimp_anaphor_number_agreement": {
+      "name": "blimp_anaphor_number_agreement",
+      "alias": "blimp_anaphor_number_agreement",
+      "sample_len": 1000,
+      "acc,none": 0.966,
+      "acc_stderr,none": 0.0057338361396953854
+    },
+    "blimp_animate_subject_passive": {
+      "name": "blimp_animate_subject_passive",
+      "alias": "blimp_animate_subject_passive",
+      "sample_len": 1000,
+      "acc,none": 0.747,
+      "acc_stderr,none": 0.013754278613587126
+    },
+    "blimp_animate_subject_trans": {
+      "name": "blimp_animate_subject_trans",
+      "alias": "blimp_animate_subject_trans",
+      "sample_len": 1000,
+      "acc,none": 0.913,
+      "acc_stderr,none": 0.008916866630745944
+    },
+    "blimp_causative": {
+      "name": "blimp_causative",
+      "alias": "blimp_causative",
+      "sample_len": 1000,
+      "acc,none": 0.663,
+      "acc_stderr,none": 0.01495508791865346
+    },
+    "blimp_complex_NP_island": {
+      "name": "blimp_complex_NP_island",
+      "alias": "blimp_complex_NP_island",
+      "sample_len": 1000,
+      "acc,none": 0.517,
+      "acc_stderr,none": 0.015810153729833274
+    },
+    "blimp_coordinate_structure_constraint_complex_left_branch": {
+      "name": "blimp_coordinate_structure_constraint_complex_left_branch",
+      "alias": "blimp_coordinate_structure_constraint_complex_left_branch",
+      "sample_len": 1000,
+      "acc,none": 0.48,
+      "acc_stderr,none": 0.015806639423035177
+    },
+    "blimp_coordinate_structure_constraint_object_extraction": {
+      "name": "blimp_coordinate_structure_constraint_object_extraction",
+      "alias": "blimp_coordinate_structure_constraint_object_extraction",
+      "sample_len": 1000,
+      "acc,none": 0.772,
+      "acc_stderr,none": 0.013273740700804434
+    },
+    "blimp_determiner_noun_agreement_1": {
+      "name": "blimp_determiner_noun_agreement_1",
+      "alias": "blimp_determiner_noun_agreement_1",
+      "sample_len": 1000,
+      "acc,none": 0.966,
+      "acc_stderr,none": 0.0057338361396953854
+    },
+    "blimp_determiner_noun_agreement_2": {
+      "name": "blimp_determiner_noun_agreement_2",
+      "alias": "blimp_determiner_noun_agreement_2",
+      "sample_len": 1000,
+      "acc,none": 0.958,
+      "acc_stderr,none": 0.006346359293033784
+    },
+    "blimp_determiner_noun_agreement_irregular_1": {
+      "name": "blimp_determiner_noun_agreement_irregular_1",
+      "alias": "blimp_determiner_noun_agreement_irregular_1",
+      "sample_len": 1000,
+      "acc,none": 0.872,
+      "acc_stderr,none": 0.010570133761108595
+    },
+    "blimp_determiner_noun_agreement_irregular_2": {
+      "name": "blimp_determiner_noun_agreement_irregular_2",
+      "alias": "blimp_determiner_noun_agreement_irregular_2",
+      "sample_len": 1000,
+      "acc,none": 0.927,
+      "acc_stderr,none": 0.008230354715243976
+    },
+    "blimp_determiner_noun_agreement_with_adj_2": {
+      "name": "blimp_determiner_noun_agreement_with_adj_2",
+      "alias": "blimp_determiner_noun_agreement_with_adj_2",
+      "sample_len": 1000,
+      "acc,none": 0.919,
+      "acc_stderr,none": 0.008632121032139986
+    },
+    "blimp_determiner_noun_agreement_with_adj_irregular_1": {
+      "name": "blimp_determiner_noun_agreement_with_adj_irregular_1",
+      "alias": "blimp_determiner_noun_agreement_with_adj_irregular_1",
+      "sample_len": 1000,
+      "acc,none": 0.826,
+      "acc_stderr,none": 0.011994493230973447
+    },
+    "blimp_determiner_noun_agreement_with_adj_irregular_2": {
+      "name": "blimp_determiner_noun_agreement_with_adj_irregular_2",
+      "alias": "blimp_determiner_noun_agreement_with_adj_irregular_2",
+      "sample_len": 1000,
+      "acc,none": 0.916,
+      "acc_stderr,none": 0.008776162089491113
+    },
+    "blimp_determiner_noun_agreement_with_adjective_1": {
+      "name": "blimp_determiner_noun_agreement_with_adjective_1",
+      "alias": "blimp_determiner_noun_agreement_with_adjective_1",
+      "sample_len": 1000,
+      "acc,none": 0.942,
+      "acc_stderr,none": 0.0073953154557929906
+    },
+    "blimp_distractor_agreement_relational_noun": {
+      "name": "blimp_distractor_agreement_relational_noun",
+      "alias": "blimp_distractor_agreement_relational_noun",
+      "sample_len": 1000,
+      "acc,none": 0.855,
+      "acc_stderr,none": 0.0111399775178902
+    },
+    "blimp_distractor_agreement_relative_clause": {
+      "name": "blimp_distractor_agreement_relative_clause",
+      "alias": "blimp_distractor_agreement_relative_clause",
+      "sample_len": 1000,
+      "acc,none": 0.687,
+      "acc_stderr,none": 0.014671272822977775
+    },
+    "blimp_drop_argument": {
+      "name": "blimp_drop_argument",
+      "alias": "blimp_drop_argument",
+      "sample_len": 1000,
+      "acc,none": 0.742,
+      "acc_stderr,none": 0.013842963108656601
+    },
+    "blimp_ellipsis_n_bar_1": {
+      "name": "blimp_ellipsis_n_bar_1",
+      "alias": "blimp_ellipsis_n_bar_1",
+      "sample_len": 1000,
+      "acc,none": 0.777,
+      "acc_stderr,none": 0.013169830843425608
+    },
+    "blimp_ellipsis_n_bar_2": {
+      "name": "blimp_ellipsis_n_bar_2",
+      "alias": "blimp_ellipsis_n_bar_2",
+      "sample_len": 1000,
+      "acc,none": 0.837,
+      "acc_stderr,none": 0.011686212712746913
+    },
+    "blimp_existential_there_object_raising": {
+      "name": "blimp_existential_there_object_raising",
+      "alias": "blimp_existential_there_object_raising",
+      "sample_len": 1000,
+      "acc,none": 0.786,
+      "acc_stderr,none": 0.012975838021968847
+    },
+    "blimp_existential_there_quantifiers_1": {
+      "name": "blimp_existential_there_quantifiers_1",
+      "alias": "blimp_existential_there_quantifiers_1",
+      "sample_len": 1000,
+      "acc,none": 0.962,
+      "acc_stderr,none": 0.0060491811505849775
+    },
+    "blimp_existential_there_quantifiers_2": {
+      "name": "blimp_existential_there_quantifiers_2",
+      "alias": "blimp_existential_there_quantifiers_2",
+      "sample_len": 1000,
+      "acc,none": 0.229,
+      "acc_stderr,none": 0.013294199326613684
+    },
+    "blimp_existential_there_subject_raising": {
+      "name": "blimp_existential_there_subject_raising",
+      "alias": "blimp_existential_there_subject_raising",
+      "sample_len": 1000,
+      "acc,none": 0.87,
+      "acc_stderr,none": 0.010640169792499236
+    },
+    "blimp_expletive_it_object_raising": {
+      "name": "blimp_expletive_it_object_raising",
+      "alias": "blimp_expletive_it_object_raising",
+      "sample_len": 1000,
+      "acc,none": 0.765,
+      "acc_stderr,none": 0.013414729030247124
+    },
+    "blimp_inchoative": {
+      "name": "blimp_inchoative",
+      "alias": "blimp_inchoative",
+      "sample_len": 1000,
+      "acc,none": 0.657,
+      "acc_stderr,none": 0.015019206922357005
+    },
+    "blimp_intransitive": {
+      "name": "blimp_intransitive",
+      "alias": "blimp_intransitive",
+      "sample_len": 1000,
+      "acc,none": 0.72,
+      "acc_stderr,none": 0.014205696104091548
+    },
+    "blimp_irregular_past_participle_adjectives": {
+      "name": "blimp_irregular_past_participle_adjectives",
+      "alias": "blimp_irregular_past_participle_adjectives",
+      "sample_len": 1000,
+      "acc,none": 0.901,
+      "acc_stderr,none": 0.009449248027662746
+    },
+    "blimp_irregular_past_participle_verbs": {
+      "name": "blimp_irregular_past_participle_verbs",
+      "alias": "blimp_irregular_past_participle_verbs",
+      "sample_len": 1000,
+      "acc,none": 0.946,
+      "acc_stderr,none": 0.007150883521295473
+    },
+    "blimp_irregular_plural_subject_verb_agreement_1": {
+      "name": "blimp_irregular_plural_subject_verb_agreement_1",
+      "alias": "blimp_irregular_plural_subject_verb_agreement_1",
+      "sample_len": 1000,
+      "acc,none": 0.882,
+      "acc_stderr,none": 0.010206869264381718
+    },
+    "blimp_irregular_plural_subject_verb_agreement_2": {
+      "name": "blimp_irregular_plural_subject_verb_agreement_2",
+      "alias": "blimp_irregular_plural_subject_verb_agreement_2",
+      "sample_len": 1000,
+      "acc,none": 0.877,
+      "acc_stderr,none": 0.010391293421849803
+    },
+    "blimp_left_branch_island_echo_question": {
+      "name": "blimp_left_branch_island_echo_question",
+      "alias": "blimp_left_branch_island_echo_question",
+      "sample_len": 1000,
+      "acc,none": 0.526,
+      "acc_stderr,none": 0.015797897758042797
+    },
+    "blimp_left_branch_island_simple_question": {
+      "name": "blimp_left_branch_island_simple_question",
+      "alias": "blimp_left_branch_island_simple_question",
+      "sample_len": 1000,
+      "acc,none": 0.584,
+      "acc_stderr,none": 0.015594460144140522
+    },
+    "blimp_matrix_question_npi_licensor_present": {
+      "name": "blimp_matrix_question_npi_licensor_present",
+      "alias": "blimp_matrix_question_npi_licensor_present",
+      "sample_len": 1000,
+      "acc,none": 0.458,
+      "acc_stderr,none": 0.015763390640483554
+    },
+    "blimp_npi_present_1": {
+      "name": "blimp_npi_present_1",
+      "alias": "blimp_npi_present_1",
+      "sample_len": 1000,
+      "acc,none": 0.536,
+      "acc_stderr,none": 0.015778243024904673
+    },
+    "blimp_npi_present_2": {
+      "name": "blimp_npi_present_2",
+      "alias": "blimp_npi_present_2",
+      "sample_len": 1000,
+      "acc,none": 0.554,
+      "acc_stderr,none": 0.01572677116675039
+    },
+    "blimp_only_npi_licensor_present": {
+      "name": "blimp_only_npi_licensor_present",
+      "alias": "blimp_only_npi_licensor_present",
+      "sample_len": 1000,
+      "acc,none": 0.964,
+      "acc_stderr,none": 0.00589395781616553
+    },
+    "blimp_only_npi_scope": {
+      "name": "blimp_only_npi_scope",
+      "alias": "blimp_only_npi_scope",
+      "sample_len": 1000,
+      "acc,none": 0.542,
+      "acc_stderr,none": 0.015763390640483554
+    },
+    "blimp_passive_1": {
+      "name": "blimp_passive_1",
+      "alias": "blimp_passive_1",
+      "sample_len": 1000,
+      "acc,none": 0.879,
+      "acc_stderr,none": 0.010318210380946179
+    },
+    "blimp_passive_2": {
+      "name": "blimp_passive_2",
+      "alias": "blimp_passive_2",
+      "sample_len": 1000,
+      "acc,none": 0.894,
+      "acc_stderr,none": 0.00973955126578524
+    },
+    "blimp_principle_A_c_command": {
+      "name": "blimp_principle_A_c_command",
+      "alias": "blimp_principle_A_c_command",
+      "sample_len": 1000,
+      "acc,none": 0.634,
+      "acc_stderr,none": 0.0152406127264056
+    },
+    "blimp_principle_A_case_1": {
+      "name": "blimp_principle_A_case_1",
+      "alias": "blimp_principle_A_case_1",
+      "sample_len": 1000,
+      "acc,none": 1.0,
+      "acc_stderr,none": 0.0
+    },
+    "blimp_principle_A_case_2": {
+      "name": "blimp_principle_A_case_2",
+      "alias": "blimp_principle_A_case_2",
+      "sample_len": 1000,
+      "acc,none": 0.925,
+      "acc_stderr,none": 0.008333333333333333
+    },
+    "blimp_principle_A_domain_1": {
+      "name": "blimp_principle_A_domain_1",
+      "alias": "blimp_principle_A_domain_1",
+      "sample_len": 1000,
+      "acc,none": 0.904,
+      "acc_stderr,none": 0.009320454434783257
+    },
+    "blimp_principle_A_domain_2": {
+      "name": "blimp_principle_A_domain_2",
+      "alias": "blimp_principle_A_domain_2",
+      "sample_len": 1000,
+      "acc,none": 0.735,
+      "acc_stderr,none": 0.013963164754810064
+    },
+    "blimp_principle_A_domain_3": {
+      "name": "blimp_principle_A_domain_3",
+      "alias": "blimp_principle_A_domain_3",
+      "sample_len": 1000,
+      "acc,none": 0.615,
+      "acc_stderr,none": 0.015395194445410945
+    },
+    "blimp_principle_A_reconstruction": {
+      "name": "blimp_principle_A_reconstruction",
+      "alias": "blimp_principle_A_reconstruction",
+      "sample_len": 1000,
+      "acc,none": 0.534,
+      "acc_stderr,none": 0.01578268332993769
+    },
+    "blimp_regular_plural_subject_verb_agreement_1": {
+      "name": "blimp_regular_plural_subject_verb_agreement_1",
+      "alias": "blimp_regular_plural_subject_verb_agreement_1",
+      "sample_len": 1000,
+      "acc,none": 0.891,
+      "acc_stderr,none": 0.009859828407037124
+    },
+    "blimp_regular_plural_subject_verb_agreement_2": {
+      "name": "blimp_regular_plural_subject_verb_agreement_2",
+      "alias": "blimp_regular_plural_subject_verb_agreement_2",
+      "sample_len": 1000,
+      "acc,none": 0.875,
+      "acc_stderr,none": 0.010463483381956722
+    },
+    "blimp_sentential_negation_npi_licensor_present": {
+      "name": "blimp_sentential_negation_npi_licensor_present",
+      "alias": "blimp_sentential_negation_npi_licensor_present",
+      "sample_len": 1000,
+      "acc,none": 0.996,
+      "acc_stderr,none": 0.001996994739098726
+    },
+    "blimp_sentential_negation_npi_scope": {
+      "name": "blimp_sentential_negation_npi_scope",
+      "alias": "blimp_sentential_negation_npi_scope",
+      "sample_len": 1000,
+      "acc,none": 0.509,
+      "acc_stderr,none": 0.0158167369950053
+    },
+    "blimp_sentential_subject_island": {
+      "name": "blimp_sentential_subject_island",
+      "alias": "blimp_sentential_subject_island",
+      "sample_len": 1000,
+      "acc,none": 0.514,
+      "acc_stderr,none": 0.01581309754773093
+    },
+    "blimp_superlative_quantifiers_1": {
+      "name": "blimp_superlative_quantifiers_1",
+      "alias": "blimp_superlative_quantifiers_1",
+      "sample_len": 1000,
+      "acc,none": 0.83,
+      "acc_stderr,none": 0.011884495834541793
+    },
+    "blimp_superlative_quantifiers_2": {
+      "name": "blimp_superlative_quantifiers_2",
+      "alias": "blimp_superlative_quantifiers_2",
+      "sample_len": 1000,
+      "acc,none": 0.681,
+      "acc_stderr,none": 0.014746404865473363
+    },
+    "blimp_tough_vs_raising_1": {
+      "name": "blimp_tough_vs_raising_1",
+      "alias": "blimp_tough_vs_raising_1",
+      "sample_len": 1000,
+      "acc,none": 0.542,
+      "acc_stderr,none": 0.015763390640483554
+    },
+    "blimp_tough_vs_raising_2": {
+      "name": "blimp_tough_vs_raising_2",
+      "alias": "blimp_tough_vs_raising_2",
+      "sample_len": 1000,
+      "acc,none": 0.923,
+      "acc_stderr,none": 0.00843458014024066
+    },
+    "blimp_transitive": {
+      "name": "blimp_transitive",
+      "alias": "blimp_transitive",
+      "sample_len": 1000,
+      "acc,none": 0.777,
+      "acc_stderr,none": 0.013169830843425608
+    },
+    "blimp_wh_island": {
+      "name": "blimp_wh_island",
+      "alias": "blimp_wh_island",
+      "sample_len": 1000,
+      "acc,none": 0.729,
+      "acc_stderr,none": 0.014062601350986125
+    },
+    "blimp_wh_questions_object_gap": {
+      "name": "blimp_wh_questions_object_gap",
+      "alias": "blimp_wh_questions_object_gap",
+      "sample_len": 1000,
+      "acc,none": 0.765,
+      "acc_stderr,none": 0.013414729030247124
+    },
+    "blimp_wh_questions_subject_gap": {
+      "name": "blimp_wh_questions_subject_gap",
+      "alias": "blimp_wh_questions_subject_gap",
+      "sample_len": 1000,
+      "acc,none": 0.913,
+      "acc_stderr,none": 0.008916866630745944
+    },
+    "blimp_wh_questions_subject_gap_long_distance": {
+      "name": "blimp_wh_questions_subject_gap_long_distance",
+      "alias": "blimp_wh_questions_subject_gap_long_distance",
+      "sample_len": 1000,
+      "acc,none": 0.911,
+      "acc_stderr,none": 0.009008893392651516
+    },
+    "blimp_wh_vs_that_no_gap": {
+      "name": "blimp_wh_vs_that_no_gap",
+      "alias": "blimp_wh_vs_that_no_gap",
+      "sample_len": 1000,
+      "acc,none": 0.969,
+      "acc_stderr,none": 0.005483527064679181
+    },
+    "blimp_wh_vs_that_no_gap_long_distance": {
+      "name": "blimp_wh_vs_that_no_gap_long_distance",
+      "alias": "blimp_wh_vs_that_no_gap_long_distance",
+      "sample_len": 1000,
+      "acc,none": 0.983,
+      "acc_stderr,none": 0.004089954489689063
+    },
+    "blimp_wh_vs_that_with_gap": {
+      "name": "blimp_wh_vs_that_with_gap",
+      "alias": "blimp_wh_vs_that_with_gap",
+      "sample_len": 1000,
+      "acc,none": 0.487,
+      "acc_stderr,none": 0.015813952101896636
+    },
+    "blimp_wh_vs_that_with_gap_long_distance": {
+      "name": "blimp_wh_vs_that_with_gap_long_distance",
+      "alias": "blimp_wh_vs_that_with_gap_long_distance",
+      "sample_len": 1000,
+      "acc,none": 0.15,
+      "acc_stderr,none": 0.011297239823409416
+    },
+    "blimp": {
+      "alias": "blimp",
+      "name": "blimp",
+      "sample_len": 67000,
+      "acc,none": 0.7658208955223881,
+      "acc_stderr,none": 0.001467664198474099,
+      "sample_count": {
+        "acc,none": 67000
+      }
+    }
+  }
+}

m08/config.json

@@ -0,0 +1,49 @@
+{
+  "model_id": "m08",
+  "run_name": "m08-gqa-rope-dense-adamw",
+  "data_dir": "data/fineweb-edu",
+  "vocab_size": 50257,
+  "block_size": 1024,
+  "d_model": 512,
+  "num_layers": 12,
+  "num_heads": 8,
+  "ffn_mult": 4,
+  "dropout": 0.0,
+  "bias": false,
+  "norm_type": "rmsnorm",
+  "tie_embeddings": true,
+  "attention_type": "gqa",
+  "gqa_num_kv_heads": 2,
+  "mla_latent_dim": 64,
+  "pos_type": "rope",
+  "rope_base": 10000,
+  "ffn_type": "dense",
+  "moe_num_experts": 8,
+  "moe_top_k": 2,
+  "optimizer": "adamw",
+  "lr": 0.0006,
+  "min_lr": 6e-05,
+  "weight_decay": 0.1,
+  "beta1": 0.9,
+  "beta2": 0.95,
+  "grad_clip": 1.0,
+  "muon_lr": 0.02,
+  "muon_momentum": 0.95,
+  "training_tokens": 1000000000,
+  "warmup_tokens": 6000000,
+  "lr_schedule": "cosine",
+  "batch_size": 4,
+  "grad_accum_steps": 16,
+  "device": "cuda",
+  "dtype": "bfloat16",
+  "compile": true,
+  "seed": 1337,
+  "eval_interval": 250,
+  "eval_iters": 100,
+  "log_interval": 10,
+  "checkpoint_interval": 1000,
+  "results_dir": "results",
+  "mlflow_experiment": "slm-ablation",
+  "mlflow_tracking_uri": "sqlite:///mlflow.db",
+  "mlflow_artifact_location": "./mlartifacts"
+}

m08/eval_results.json

@@ -0,0 +1,212 @@
+{
+  "model_id": "m08",
+  "checkpoint": "results/m08/final.pt",
+  "trained_step": 15258,
+  "total_params": 58774528,
+  "active_params": 58774528,
+  "flops_per_forward": 117549056.0,
+  "val_loss": 3.468865367770195,
+  "val_perplexity": 32.10029973945804,
+  "bits_per_token": 5.00451486359353,
+  "ppl_tokens_evaluated": 655360,
+  "prefill_tok_per_sec": 176358.66558708608,
+  "inference_tok_per_sec": 215.73418372595413,
+  "inference_tok_per_sec_kvcache": 210.01077375697125,
+  "kvcache_speedup": 0.9734700830895984,
+  "gen_tokens": 128,
+  "peak_infer_vram_gb": 0.366,
+  "peak_train_vram_gb": 3.993,
+  "closeness_answer_ppl": 33.412,
+  "closeness_hit_rate": 0.35,
+  "closeness_token_f1": 0.017,
+  "closeness_n": 20,
+  "closeness_items": [
+    {
+      "prompt": "The capital of France is",
+      "answer": " Paris",
+      "generated": " the capital of the French Republic",
+      "answer_nll": 3.044,
+      "answer_ppl": 20.99,
+      "hit": false,
+      "token_f1": 0.0
+    },
+    {
+      "prompt": "The capital of Japan is",
+      "answer": " Tokyo",
+      "generated": " the capital of the Republic of",
+      "answer_nll": 3.2951,
+      "answer_ppl": 26.979,
+      "hit": false,
+      "token_f1": 0.0
+    },
+    {
+      "prompt": "The largest planet in the solar system is",
+      "answer": " Jupiter",
+      "generated": " the solar system.\nThe",
+      "answer_nll": 4.1156,
+      "answer_ppl": 61.291,
+      "hit": false,
+      "token_f1": 0.0
+    },
+    {
+      "prompt": "Water is made of hydrogen and",
+      "answer": " oxygen",
+      "generated": " oxygen.\nThe hydrogen atom",
+      "answer_nll": 1.0477,
+      "answer_ppl": 2.851,
+      "hit": true,
+      "token_f1": 0.0
+    },
+    {
+      "prompt": "The opposite of hot is",
+      "answer": " cold",
+      "generated": " the same.\nThe first",
+      "answer_nll": 3.7453,
+      "answer_ppl": 42.323,
+      "hit": false,
+      "token_f1": 0.0
+    },
+    {
+      "prompt": "The opposite of up is",
+      "answer": " down",
+      "generated": " the fact that the two are",
+      "answer_nll": 8.439,
+      "answer_ppl": 4623.824,
+      "hit": false,
+      "token_f1": 0.0
+    },
+    {
+      "prompt": "Two plus two equals",
+      "answer": " four",
+      "generated": " two equal numbers of the two",
+      "answer_nll": 2.7378,
+      "answer_ppl": 15.452,
+      "hit": false,
+      "token_f1": 0.0
+    },
+    {
+      "prompt": "The sun rises in the",
+      "answer": " east",
+      "generated": " sky, and the sun is",
+      "answer_nll": 4.242,
+      "answer_ppl": 69.545,
+      "hit": false,
+      "token_f1": 0.0
+    },
+    {
+      "prompt": "The first president of the United States was George",
+      "answer": " Washington",
+      "generated": " Washington. He was the first",
+      "answer_nll": 0.3207,
+      "answer_ppl": 1.378,
+      "hit": true,
+      "token_f1": 0.0
+    },
+    {
+      "prompt": "The speed of light is approximately 300,000 kilometers per",
+      "answer": " second",
+      "generated": " second.\nThe speed of",
+      "answer_nll": 0.0622,
+      "answer_ppl": 1.064,
+      "hit": true,
+      "token_f1": 0.0
+    },
+    {
+      "prompt": "A triangle has three",
+      "answer": " sides",
+      "generated": " sides, and the triangle has",
+      "answer_nll": 1.5957,
+      "answer_ppl": 4.932,
+      "hit": true,
+      "token_f1": 0.0
+    },
+    {
+      "prompt": "The chemical symbol for gold is",
+      "answer": " Au",
+      "generated": " the gold.\nThe symbol",
+      "answer_nll": 8.9649,
+      "answer_ppl": 7823.719,
+      "hit": false,
+      "token_f1": 0.0
+    },
+    {
+      "prompt": "The freezing point of water in Celsius is",
+      "answer": " zero",
+      "generated": " about 1.5\u00b0C",
+      "answer_nll": 5.2632,
+      "answer_ppl": 193.092,
+      "hit": false,
+      "token_f1": 0.0
+    },
+    {
+      "prompt": "The author of Romeo and Juliet is William",
+      "answer": " Shakespeare",
+      "generated": " Shakespeare, the author of Romeo",
+      "answer_nll": 0.4581,
+      "answer_ppl": 1.581,
+      "hit": true,
+      "token_f1": 0.0
+    },
+    {
+      "prompt": "The Earth revolves around the",
+      "answer": " Sun",
+      "generated": " sun, and the sun is",
+      "answer_nll": 1.2541,
+      "answer_ppl": 3.505,
+      "hit": true,
+      "token_f1": 0.333
+    },
+    {
+      "prompt": "The plural of mouse is",
+      "answer": " mice",
+      "generated": " a word that is used to",
+      "answer_nll": 8.9012,
+      "answer_ppl": 7340.772,
+      "hit": false,
+      "token_f1": 0.0
+    },
+    {
+      "prompt": "Photosynthesis occurs in the",
+      "answer": " leaves",
+      "generated": " soil, and it is a",
+      "answer_nll": 4.1607,
+      "answer_ppl": 64.115,
+      "hit": false,
+      "token_f1": 0.0
+    },
+    {
+      "prompt": "The human heart pumps",
+      "answer": " blood",
+      "generated": " through the heart.\nThe",
+      "answer_nll": 3.1444,
+      "answer_ppl": 23.205,
+      "hit": false,
+      "token_f1": 0.0
+    },
+    {
+      "prompt": "The currency used in the United States is the",
+      "answer": " dollar",
+      "generated": " US dollar.\nThe US",
+      "answer_nll": 2.9611,
+      "answer_ppl": 19.32,
+      "hit": true,
+      "token_f1": 0.0
+    },
+    {
+      "prompt": "The tallest animal in the world is the",
+      "answer": " giraffe",
+      "generated": " giant tortoise.\nThe",
+      "answer_nll": 2.4257,
+      "answer_ppl": 11.311,
+      "hit": false,
+      "token_f1": 0.0
+    }
+  ],
+  "generation_samples": {
+    "The most important thing about sleep is": "The most important thing about sleep is that it\u2019s easier to sleep when you\u2019re hungry than when you\u2019re thirsty.\nIt\u2019s important to focus on the most basic sleep patterns you can expect. If you\u2019re hungry, you\u2019ve got the best night\u2019s rest,",
+    "Scientists have discovered that": "Scientists have discovered that a large asteroid that\u2019s about the size of a football.\nThe asteroid hit the Earth in 1999. The asteroid is about the size of Alaska that\u2019s hit the globe in 1999.\nBut asteroids also hit our planet, and it\u2019s likely that the asteroid has killed",
+    "In the history of mathematics,": "In the history of mathematics, the ancient Greek philosopher and physician, Theologus\u2014the Greek philosopher and physician\u2014is described as a human being. The term \u201cGreek\u201d is widely used in the 19th Century, but the word \u201cGreek\u201d is a derivative of the Greek word \u201cepicos",
+    "Python code for prime number": "Python code for prime number systems, for example. However, the main difference between two solutions is their size. In the case where two solutions are identical, the first solution is called the highest number of prime numbers, so the second solution is called the worst number of prime numbers.\nThe numbers in the prime number list,",
+    "The best way to learn a new language is": "The best way to learn a new language is to learn about it.\nThe first step in learning new language is to learn something new.\nThe first step to learning new language is to learn new words.\n\"Reading is not the only thing that can teach us anything. We have to learn anything we can learn. You can learn about"
+  }
+}

m08/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:984e2231d35f9184f4d8c47f643bc8f60ee69e86a62961b0d57b53d1df050215
+size 235107624